[gmx-users] Comparison of PMF profiles in different states of a protein

[gmx user1]

Hi

I am studying the movement of a molecule in two different states of a
protein. To do so, I want to calculate a PMF profile in each state using
umbrella sampling. My reaction coordinate is the distance between the
center-of-mass of a fixed residue and the center-of-mass of the molecule.
My goal is to compare both profiles.

Although I am using an internal coordinate of the system, I need to make
the profiles comparable and to eliminate fluctuations of the reference
residue (which is already in a rigid zone of the protein). To that end, I
am thinking to adopt a similar procedure to the one used by Hub et al (
https://pubs.acs.org/doi/abs/10.1021/ja102133x), in which I plan to align
the trajectories of all windows to a global average structure and I
calculate the difference between the coordinates of the reference residue
on the trajectory and of the average structure. Only after, I calculate the
free energy with the modified CV.

Can anyone give me an opinion whether this procedure makes any sense?

Furthermore, how can I ensure that both profiles have the same energy
zero?  If I use WHAM to calculate the profiles, even if I supply initial
conformations with the same CV value to both states (because g_wham zeroes
the profiles at the first supplied conformation), due to WHAM reweigthing
the weights of the conformations in each state will be different? Am I
correct? In that case would it be better to use umbrella integration?

Any help is appreciated.

Thanks in advance.